Lateral gene transfer is a nearly ubiquitous influence on bacterial genome evolution, introducing a wide palette of novel genetic material the can facilitate adaptation; this is especially true in pathogens, where genomic islands have introduced metabolic functions allowing the adoption of the pathogenic lifestyle. While a large number of laterally transferred genes are recognized in recipient genomes, information regarding potential donor pools is limited. Since lateral gene transfer is such a powerful force for change, the nature of donor taxa - including their phylogenetic relationships and metabolic capabilities - provides a context for understanding the evolution of extant lineages, as well as insight into their evolutionary potential. Critically, there are several theoretical and empirical arguments for why donor taxa may not be chosen at random from all bacteria, but rather may be more closely related to the recipients. In this way, bacterial clades may form under the influence of gene transfer, and may not simply reflect the common ancestry of shared genes. We propose to identify the likely donor taxa for the recently-acquired gene pools of major bacterial clades. First, an artificial genome framework will be used to develop a robust method for utilizing and combining parametric methods for alien gene detection. Second, we will develop both gene cluster and genome segmentation methods for the identification of alien genes, using artificial genomes and parametric guidelines established above. These methods will identify groups of atypical genes that are similar to one another. The idea is that atypical genes with similar properties may have originated from the same source. Using these clusters of genes, we will ascertain their likely ancestry by extrapolating phylogenetic methods to infer the donor taxa for genes without homologues in the database. The assignment of donor taxa will be verified by a novel analysis of replication bias patterns, allowing us to correlate strand-bias patterns of alien genes with the signatures of the genomes in their donor's clades, which can be phylogenetically quite broad. This framework will allow us to test the hypothesis that gene transfer occurs more often between more closely related lineages. [unreadable] [unreadable] [unreadable]